International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 06 Issue: 01 | Jan 2019
p-ISSN: 2395-0072
www.irjet.net
DATA MINING ALGORITHM IN DISTRIBUTED DATABASE
Vivek Gohil1, Ketan Gharge2, Ritesh Ghorui3, Silviya Dmonte4
1,2,3Student,
4Assistant
Dept of Computer Engineering, Universal College, Maharashtra, India
Professor, Dept of Computer Engineering, Universal College, Maharashtra, India
-----------------------------------------------------------------------------***---------------------------------------------------------------------------Abstract – Data mining is a form of business intelligence and
specific inquiry. Basically, it inverts the experimental system,
data analysis. It is the process of analysing data to draw
beginning from information and moving towards
useful conclusions or predictions from it. It’s a technique
speculations as opposed to taking after the conventional
frequently adopted by large-scale ecommerce businesses to
request (Berry, et al. 2000).
aid with marketing and product development. Because of the
nature of the internet, ecommerce businesses will obtain a
While data mining and learning disclosure in databases
lot of data about their customers, or their prospective
(KDD) are as regularly as conceivable viewed as
customers. Data is obtained whenever a purchase is made,
proportionate words, information mining is really piece of
an account is created, or a page view is made. This raw data
the learning disclosure process (Zaiane 1999). The
is obtained from the Distributed database. A utilization of a
accompanying shows information mining as a venture in
calculations to dispersed databases isn't successful, as it
learning disclosure process.
requires a lot of correspondence overhead. In our
investigation, a proficient calculation, EDMA (Effective Data
Information extraction methodology separates valuable
Mining Algorithm), is proposed. It minimizes the number of
subsets of information for mining. Objectives of the
candidate sets and exchange messages by local and global
extraction procedure are, recognizing concerned data in the
pruning.
database and transforming the database into some suitable
structure to investigation by the information mining
Key Words: Apriori algorithm, Association rules, parallel
calculations.
and distributed data mining.
Information arrangement is a standout amongst the most
1. INTRODUCTION
essential ventures in the information mining procedure. Vast
database frameworks by and large contain mistakes in the
Information mining is the act of looking at huge previous
put away information. Inspect the information for blunders,
databases keeping in mind the end goal to produce new data
frameworks and missing qualities to the nature of the
There are different kinds of data mining techniques are
information. This is the most drawn out and most critical
available. Classification, Clustering, Association Rule and
venture in the information planning methodology. Strength
Neural Network Weka are some of the most significant
is an imperative property for the information mining
frameworks. Thus, a few procedures are accustomed to
techniques in data mining.
figuring out how to information in this methodology.
In business world today, Online Shopping has become very
Information cleaning can be connected to evacuate
popular means of shopping. It first started with just simple
commotion and irregularity in the information. There are
book store but now everything is available online at very
numerous purposes behind uproarious and deficient
reasonable prices. Thus with the help of Data Mining it’s
information. Some of strategies are accustomed to filling in
possible to increase the sales of the products. In this project
the missing qualities. The most acclaimed one is the relapse
data mining is used to analyze the transaction data and
systems for information cleaning. Information coordination
group the products which are most frequently purchased
combines information from numerous sources. These
together. When customer searches for a product, all the
sources may incorporate numerous databases, information
products which were frequently purchased together with the
blocks, or level documents.
searched product will be displayed to him along with the
product he searched for. It is natural when related products
The principle issue of the incorporation is information
are displayed together, a person`s desire increases to
confliction. Information change operations utilized for
purchase both of them. E.g. If we keep an offer on Mobile, it`s
normalizations and conglomeration. Information are
Screen protector and a back cover together then it is more
changed or incorporated into structure suitable for mining.
probable that customer will purchase all three instead of just
Information change methodology incorporates smoothing,
mobile.
speculation, standardization and collection methods.
Information diminishment operation utilized for lessen the
Information mining varies from other exploration system in
information measure by utilizing one of the information
that it is proposed to deal with information without
collection, measurement decrease or information
beginning from a specific theory, presumption or even a
examination strategies. Information diminishment
© 2018, IRJET
|
Impact Factor value: 7.211
|
ISO 9001:2008 Certified Journal
|
Page 189
International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 06 Issue: 01 | Jan 2019
p-ISSN: 2395-0072
www.irjet.net
techniques can be accustomed to minimizing representation
of the information, while diminishing the loss of data
substance.
model building accessible to business users are also
evaluated An unmistakable manual for web based business
organizations sitting on colossal volume of information to
effortlessly control the information for business change
which consequently will put them exceedingly focused
among their rivals is additionally given in this paper. [2]
The greater part of the crude information are made and
cleaned in the past steps. Hence, information is arranged for
the information mining stage. Arranged information may
contain numerous credits and we need to choose a subset of
the qualities for utilizing as a part of information mining
methodology.
A. Prodromidis, P. Chan, and S. Stolfo. In their work on
Parallal distributed data mining systems: Issues and
approaches in AAAI/MIT Press, 2001.Association Rule
Learning is a general technique used to discover associations
amongst numerous variables. It is often used by grocery
stores, retailers, and anyone with a bulky transactional
database [7].
A Data mining calculation takes information as data and
produces yield as models or examples. In this step a canny
strategies are connected keeping in mind the end goal to
concentrate information designs. Visualization, order,
grouping, relapse or affiliation calculations are utilized for
diverse issue. There are numerous algorithmic ways to deal
with separating helpful data from information.
3. Proposed System
Since Mining is the crucial and very important in every
sector of business it cannot be ruled out.
To make this system faster & efficient we divide or split
larger chuck of operational data on multiple physical
database
2. Literature Review
The following research articles are selected for review,
keeping in mind the traditional and conventional approach
of an efficient association rule mining algorithm in
distributed database.
Neha Saxena, Rakhi Arora, Ranjana Sikarwar and Ashika
Gupta in their study of An Efficient Approach of Association
Rule Mining on Distributed Database Algorithm journal
Applications requiring huge data processing have two main
problems, one a massive storage and its supervision and
next processing time, when the quantity of data increases.
Distributed databases determine the first trouble to a huge
amount but second problem increase. Since, current stage is
of networking and communication and community are
involved in maintenance huge data on networks, therefore,
researchers are propose a range of novel algorithms to raise
the throughput of resulted data over distributed databases.
Within our research, we are proposing an novel algorithm to
process large quantity of data at the a variety of servers and
collect the processed data on customer machine as much as
necessary [1]
Figure 1. System Architecture
Following are the modules of this system:
3.1. Authentication,Authorization ,Registration Module:

3.2. Admin :
Mustapha Ismail, Mohammed Mansur Ibrahim, Zayyan
Mahmoud Sanusi, Muesser Nat in their Information Mining
in Electronic Business: The fundamental point of this paper
is to audit the utilization of information mining in web based
business by concentrating on organized and unstructured
information gathered exhaustive different assets and
distributed computing administrations keeping in mind the
end goal to legitimize the significance of information mining.
Also, this examination assesses certain difficulties of
information mining like creepy crawly recognizable proof,
information changes and influencing information to
demonstrate fathomable to business clients. Other
challenges which are supporting the slow changing
dimensions of data, making the data transformation and
© 2018, IRJET
|
Impact Factor value: 7.211
This will be secured prototype means every user
has to sign up for using the prototype. Based on the
role they will redirected to their corresponding
home page.



Admin can add the product.
Admin can view the details of the users.
Admin can add the new category of the product.
3.3 Users :




|
Search for the product.
User can view own details & can make changes.
Can add the product to cart.
Can initiate new Transaction.
ISO 9001:2008 Certified Journal
|
Page 190
International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 06 Issue: 01 | Jan 2019
p-ISSN: 2395-0072
www.irjet.net
3. 4. Mining Module :






There will be parallel mining performed at each
distributed node on operational data.
Mining algorithm will take past user transaction as
input and generated corresponding output.
This will generate various resultant set containing
relative between products.
If above mentioned problems are considered, it is possible to
improve the sales by increasing customer`s desire of
shopping.



3.5. Fragmentation :



Accessibility,
Quality,
Availability,
It is technique of splitiing data on various
distributed node.
We have chosen Hybrid fragmentation approach.
We will split the product based on category.This
will fragment each product table at every node but
with different product data.


To analyze the transactions.
To shortlist products.
To examine the product which customer searches
most frequently.
To give the suggestions on the basis of finding of the
products.
To suggest a good relative product along with the
searched product.
3. 7. Query optimization :
 Query is run on subset of data resulting faster
retrivation hence waiting time is reduced.
With the help of Data Mining it’s possible to increase the
sales of the products. In this project data mining is used to
analyze the transaction data and group the products which
are most frequently purchased together. When customer
searches for a product, all the products which were
frequently purchased together with the searched product
will be displayed to him along with the product he searched
for. It is natural when related products are displayed
together, a person`s desire increases to purchase both of
them. E.g. If we keep an offer on Mobile, it`s Screen protector
and a back cover together then it is more probable that
customer will purchase all three instead of just mobile.
3. 8. Technical feasibility :
5. CONCLUSION
3.6. Parallel processing :







When user search for product .search on separate
node takes place simultaneously.
This improves performance.
Also mining algorithm on each node will run
parallely so that waiting time will be reduced.
Analyzing the previous records in order to boost up the
current business with help of identifiable pattern is a good
strategy for increase in business which ultimately leads to
increased profit. The aim of data mining is to extract
knowledge from information stored in database and
generate clear and understandable description of patterns.
The main aim of this paper is to predict possible
combinations of products which are more likely to be
purchased together with the help of Apriori algorithm for
data mining tool. Then this algorithm was implemented
using Association data mining technique. This algorithm was
made to run after scheduled time so that the time taken by
this algorithm do not affect the searching time of user as it
would have if it was implemented such that it executes at
time when user hits ‘Search’ button. This algorithm shortlists
the products in to on basis of how frequently they are
purchased together and helps us in displaying those grouped
products with any one of the searched product so that user
gets attracted towards the combos or such offers. This
algorithm can also be used in other systems like E Learning.
In that same thing can be applied for analyzing most
frequently accessed tutorials by a user and group tutorials
into a set consisting of tutorials of related subjects. Eg if a
searches for Core Java Tutorial, then by analyzing history of
other users, we can create a set of tutorials related to java,
We have chosen Java-8 as implementation language.
My-SQL 5.6 as Relational Database.
For powerful front end we have chosen JSP.
To host application we have chosen Apache Tomcat
7.0
4. Discussion
Electronic business, or e-business, is the use of data and
correspondence advancements (ICT) in help of the
considerable number of exercises of business. Trade
constitutes the trading of items and administrations between
organizations, gatherings and people and can be viewed as
one of the fundamental exercises of any business Electronic
trade centers around the utilization of ICT to empower the
outside exercises and connections of the business with
people, gatherings and different organizations or e business
alludes to business with help of web i.e. working with the
assistance of web organize.
Mostly, customers are attracted to such software, tools, etc.
which are easy to use and user friendly. Thus it would be
beneficial to develop a Shopping site which is more userfriendly and with a good User Interface. Following points of
improvements can be considered while making such sites.
© 2018, IRJET
|
Impact Factor value: 7.211
|
ISO 9001:2008 Certified Journal
|
Page 191
International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 06 Issue: 01 | Jan 2019
p-ISSN: 2395-0072
www.irjet.net
i.e. say Applet or AWT and suggest them to other users. This
way we can make it user friendly.
REFERENCES
[1] A. W. Jilani,” Usage and Effectiveness of E-Commerce Tool
amongBusiness-To-Consumer”, 2017.
[2] Neha Saxena, Rakhi Arora, Ranjana Sikarwar and Ashika
Gupta, “An Efficient Approach of Association Rule Mining
on Distributed Database”, Volume 3, Number 4 (2013).
[3] Mustapha Ismail, Mohammed Mansur Ibrahim, Zayyan
Mahmoud Sanusi, Muesser Nat,”Data Mining in Electronic
Commerce”, 2015.
[4]Miva, M. and Miva, B.The History of Ecommerce: How Did
It All Begin?http://www.miva.com/blog/the-history-Of
ecommerce-how -didit-all-begin. 2011
[5] A. Prodromidis, P. Chan, and S. Stolfo,”Parellal,distributed
datamining systems”,2001.
© 2018, IRJET
|
Impact Factor value: 7.211
|
ISO 9001:2008 Certified Journal
|
Page 192